The Silicon Valley startup K2 Cyber Security is trying to make a name for itself by saying that it has a product that will stop any zero-day attack. It claims it can do this in the cloud by monitoring how an app performs and comparing it to how it thinks the app should perform.
The company says that, "K2 has developed Optimized CFI, the first deterministic solution that promises to enable no false positive, highly precise software-based CFI to enterprises, with easy operationalization: no need for source code for instrumentation, expensive re-compiles or hardware forklifts to implement. K2 Prevent works via an agent that knows the legitimate execution paths of your application infrastructure and instantly alerts when an exploit or web attack occurs. Unlike most security approaches, K2 does not depend on prior behavioral or signature-based knowledge of any attack, enabling the solution to be effective against zero-day attack."
As claims go, that's pretty out there.
In the past, CFI has only been feasible with select hardware and OS implementations. Re-compiles of applications and hardware upgrades may have been needed to get it to function.
Also, CFI has had operational trade-offs in the software like instrumentation, heavy infrastructure or high additional CPU overhead requirements. This has made adoption of it by most enterprise IT environments problematic.
The company describes its process simply. Agents are deployed on physical or virtual servers and VMs via an installer and as Kubernetes minion nodes as pods. The K2 agent "automatically identifies applications and protect[s] them from attacks." An inconsistency in execution of an application due to an attack should generate a security alert in real time which can then be viewed in user's own GUI or in SIEM solution.
K2 says that it has the process all worked out. But so far, no operational details such as tests or code examples have been provided to the public by the company to verify its claims. The folk at K2 say they have patents pending on the optimizations, so that may be a factor here.
Ankit Anubhav, the principal security researcher of Russian-based NewSkySecurity, is also somewhat confused by K2's lack of transparency.
He told Security Now: "It's an interesting approach to use the whitelisting so anything besides normal behavior should be tracked and can be helpful for catching an unknown attack or zero day. That being said, it's too abstract and top level for me to comment on it as the technical details, real-world use cases or code isn't shared."
K2 has claimed much, and now needs to show potential customers what it has under the hood for them.
— Larry Loeb has written for many of the last century's major "dead tree" computer magazines, having been, among other things, a consulting editor for BYTE magazine and senior editor for the launch of WebWeek.